research communications

398 https://doi.org/10.1107/S2053230X17008093 Acta Cryst. (2017). F73, 398–403

Received 31 March 2017

Accepted 31 May 2017

Edited by A. Nakagawa, Osaka University, Japan

‡ Current address: Institute for Molecular

Science, National Institute of Natural Science,

5-1 Higashiyama, Myodaiji, Okazaki, Aichi

444-8787, Japan.

§ Current address: EM Utilization Center,

Human Spaceflight Technology Directorate,

Japan Aerospace Exploration Agency (JAXA),

2-1-1 Sengen, Tsukuba-shi, Ibaraki-ken, 305-

8505 Japan,

Keywords: cellulases; Phanerochaete

chrysosporium; cellobiohydrolase; biomass

utilization; carbohydrate-active enzymes.

PDB references: PcCel6A, 5xcy; complex with

cellobiose, 5xcz

Supporting information: this article has

supporting information at journals.iucr.org/f

Crystal structure of a family 6 cellobiohydrolasefrom the basidiomycete Phanerochaetechrysosporium

Mikako Tachioka,a Akihiko Nakamura,a‡ Takuya Ishida,a§ Kiyohiko Igarashia,b and

Masahiro Samejimaa*

aDepartment of Biomaterial Sciences, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1

Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan, and bVTT Technical Research Centre of Finland, PO Box 1000, Tietotie 2,

Espoo FI-02044 VTT, Finland. *Correspondence e-mail: amsam@mail.ecc.u-tokyo.ac.jp

Cellobiohydrolases belonging to glycoside hydrolase family 6 (CBH II, Cel6A)

play key roles in the hydrolysis of crystalline cellulose. CBH II from the white-

rot fungus Phanerochaete chrysosporium (PcCel6A) consists of a catalytic

domain (CD) and a carbohydrate-binding module connected by a linker

peptide, like other known fungal cellobiohydrolases. In the present study, the

CD of PcCel6A was crystallized without ligands, and p-nitrophenyl �-d-cello-

trioside (pNPG3) was soaked into the crystals. The determined structures of the

ligand-free and pNPG3-soaked crystals revealed that binding of cellobiose at

substrate subsites +1 and +2 induces a conformational change of the N-terminal

and C-terminal loops, switching the tunnel-shaped active site from the open to

the closed form.

1. Introduction

Cellulose is the most abundant biopolymer on Earth, and

is a potential alternative resource to fossil-based fuels and

chemicals. It consists of �-1,4-linked d-glucose units, and is

mainly found in plant cell walls in highly crystalline forms,

which provide mechanical strength and resistance to microbial

and chemical breakdown. Cellulose-degrading microorgan-

isms have evolved to produce an array of extracellular

enzymes that convert cellulose into soluble oligosaccharides.

Cellobiohydrolases (EC 3.2.1.91), which hydrolyze cellulose to

cellobiose (a �-1,4-linked d-glucose dimer), are key members

of their enzyme cocktails, playing a major role in the hydro-

lysis of crystalline cellulose.

Glycoside hydrolase (GH) family 6, a class of enzymes

defined by amino-acid sequence similarities in the CAZy

database (Carbohydrate-Active enZymes; http://www.cazy.org),

includes cellobiohydrolases and endoglucanases (EC 3.2.1.4).

Structural studies have revealed that cellobiohydrolases of this

family have tunnel-like catalytic sites covered by flexible

loops, which are not necessarily conserved in endoglucanases

(Rouvinen et al., 1990; Spezio et al., 1993). This family is also

known for its unusual ‘Grotthus-type’ inverting mechanism:

the enzymes appear to lack a general base residue that would

activate a catalytic water molecule for direct nucleophilic

attack; instead, observations of ordered water molecules near

the active site led to the proposal that an additional water

molecule is involved in hydrolysis, serving to bridge between

the catalytic water molecule and the proton-accepting residue

(Koivula et al., 2002).

The GH family 6 cellobiohydrolase from the basidiomycete

P. chrysosporium (PcCel6A) consists of a catalytic domain

ISSN 2053-230X

(CD) and an N-terminal carbohydrate-binding module (CBM)

connected by a linker region (Tempelaars et al., 1994); this

modular structure is typical of fungal GH family 6 cellobio-

hydrolases (Mertz et al., 2005). In a previous study, we cloned

and recombinantly expressed PcCel6A and found that the

hydrolysis rate of PcCel6A was greatly accelerated by a

polymorphic change of crystalline cellulose (Igarashi et al.,

2012). PcCel6A has been the target of protein engineering

aimed at improving its thermostability (Heinzelman et al.,

2009; Ito et al., 2013), and we have also recently applied

random mutagenesis to this enzyme (Tachioka et al., 2016).

Here, we report the structure of the CD of PcCel6A in its apo

form and in complex with the ligand cellobiose.

2. Materials and methods

2.1. Expression and purification

Cloning of the cel6A gene from P. chrysosporium and the

construction of the pPICZ�/cel6A vector have been described

in a previous report (Igarashi et al., 2012). Deletion of the

N-terminal CBM-linker region (residues 1–81) was performed

by a combination of inverse PCR and DpnI treatment. The

recombinant protein was produced in P. pastoris using a 5 l jar

fermenter, as reported previously (Igarashi et al., 2012), and

was then ultrafiltered and concentrated using a Kvick Lab

Cassette 100 kDa and 5 kDa (GE Healthcare, USA). The

protein solution in 20 mM sodium acetate buffer pH 5.0

containing 1 M ammonium sulfate was applied onto a Phenyl

Toyopearl 650S column (Tosoh Corporation, Japan) equili-

brated with the same buffer. The protein was eluted with a

reverse gradient to 20 mM sodium acetate buffer pH 5.0 and

was then analysed by SDS–PAGE. Doublet bands were seen at

approximately 37 kDa, which might reflect a difference in

glycosylation. Therefore, the fractions with lower molecular

weight were collected. These fractions were equilibrated

against 20 mM Tris–HCl buffer pH 8.0 and applied onto a

DEAE Toyopearl 650S column (Tosoh Corporation) equili-

brated with the same buffer. The protein was eluted from the

column with a linear gradient from 0 to 20 mM NaCl. The

purified protein was dialyzed into 5 mM Tris–HCl buffer pH

7.5 containing 100 mM NaCl.

2.2. Crystallization, data collection and structure solution

Crystallization was performed by the sitting-drop vapour-

diffusion method. The drops were formed by mixing 1 ml

20 mg ml�1 protein solution with the same volume of a

reservoir solution composed of 20%(w/v) polyethylene glycol

3350, 200 mM calcium acetate, 50 mM acetate buffer pH 5.0,

10%(w/v) 2-methyl-2,4-pentanediol. To introduce the ligand,

p-nitrophenyl �-d-cellotrioside (pNPG3) powder was

dissolved in mother liquor and a crystal was incubated in this

solution for 10 h prior to data collection. X-ray diffraction

data sets were collected using synchrotron radiation on

beamlines BL5A and BL17A of the Photon Factory, High

Energy Accelerator Research Organization (KEK), Tsukuba,

Japan. The data sets were processed and scaled using the

HKL-2000 suite (Otwinowski & Minor, 1997). The sequence

of PcCel6A was submitted to the Phyre2 web server (Kelley &

Sternberg, 2009; http://www.sbg.bio.ic.ac.uk/phyre2) to obtain

a suitable search model for molecular replacement with

MOLREP, an auto-MR function in the CCP4 suite (Winn et

al., 2011). Manual model rebuilding and refinement were

performed using Coot (Emsley et al., 2010) and PHENIX

(Adams et al., 2010). Data-collection and refinement statistics

are shown in Table 1. Most of the molecular-graphics images

were prepared using PyMOL (v.1.7; Schrodinger). The rela-

tive B factor was calculated by dividing the average B factor of

each residue by that of the whole protein: 11.6 A2 for the apo

structure and 20.0 A2 for the PcCel6A–cellobiose structure.

The apo and cellobiose-bound structures were superposed

using the phenix.superpose_maps tool, and the r.m.s.d. values

between all C�-atom pairs and the relative B-factor values

were visualized using UCSF Chimera (http://www.cgl.ucsf.edu/

chimera; Pettersen et al., 2004). Hydrophobicity was visualized

using the color_h script in PyMOL.

3. Results and discussion

3.1. Overall structure of the PcCel6A catalytic domain

X-ray data sets were collected from a PcCel6A crystal and a

pNPG3-soaked crystal to 1.2 and 2.1 A resolution, respec-

tively. Both crystals belonged to space group P212121 with one

molecule in the asymmetric unit, and the structures were

research communications

Acta Cryst. (2017). F73, 398–403 Tachioka et al. � Family 6 cellobiohydrolase 399

Table 1X-ray data-collection and refinement statistics.

Values in parentheses are for the highest resolution shell.

PcCel6A PcCel6A–cellobiose

Data collectionBeamline BL5A BL17AWavelength (A) 1.00000 0.98000Space group P212121 P212121

Unit-cell parametersa (A) 54.7 54.5b (A) 67.2 67.0c (A) 89.1 85.1

Resolution (A) 50.00–1.20 (1.22–1.20) 50.00–2.10 (2.14–2.10)Total reflections 1014578 129843Unique reflections 103410 19086Completeness (%) 99.9 (99.2) 99.5 (99.0)Multiplicity (%) 3.1 (2.8) 6.8 (6.8)Average I/�(I) 43.7 (5.9) 15.5 (3.1)Rmerge (%) 5.5 (28.8) 7.2 (48.4)Mosaicity range (�) 0.26–0.32 0.79–1.27

RefinementResolution (A) 28.6–1.20 (1.24–1.20) 45.9–2.1 (2.18–2.10)Rwork (%) 13.8 (14.9) 17.7 (21.9)Rfree (%) 15.9 (19.6) 23.5 (27.6)No. of reflections 103210 (10094) 18676 (1806)No. of atoms 3515 2950R.m.s.d. from ideal values

Bond lengths (A) 0.005 0.007Bond angles (�) 0.83 0.85

Ramachandran plotFavoured regions (%) 96.9 97.0Additionally allowed (%) 3.1 3.0Outliers (%) 0 0

refined to Rwork and Rfree values of 13.8 and 15.9% and of 17.7

and 23.5%, respectively.

The CD of PcCel6A consists of a distorted seven-stranded

�/�8-barrel, like the CDs of other fungal GH family 6 cello-

biohydrolases (Fig. 1). The final structures included residues

82–439. The catalytic site is located in a tunnel enclosed by a

pair of loops (residues 174–178 and 390–425), designated here

as the N-terminal and C-terminal loops (shown in green and

blue, respectively, in Fig. 1). A structure-similarity search of

the Protein Data Bank (PDB) using the DALI server (Holm

& Rosenstrom, 2010) revealed highly homologous structures

among GH family 6 enzymes. The overall structure and ligand-

bound structure of PcCel6A showed the greatest similarity to

Cel6A from Coprinopsis cinerea (CcCel6A; Tamura et al.,

2012; PDB entry 3vog), with an r.m.s.d. of 0.7 A for 357 C�

atoms and a Z-score of 60.8, and to CcCel6A–pNPG3 (PDB

entry 3voi), with an r.m.s.d. of 0.9 A for 357 C� atoms and a

Z-score of 61.3, respectively. The sequence similarity between

PcCel6A and CcCel6A was 65%. The catalytically important

residues and two conserved disulfide bridges (Cys171–Cys230

and Cys361–Cys408) are shown in Supplementary Fig. S1,

together with sequence alignments of PcCel6A, CcCel6A,

Cel6A from Trichoderma reesei (TrCel6A) and Cel6A from

Humicola insolens (HiCel6A). The CD of PcCel6A has one

potential N-glycosylation site at Asn398 according to the

NetNglyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/),

but no electron density owing to sugars was visible in the

structures, indicating that the proteins used to prepare the

crystals were nonglycosylated.

3.2. Ligand binding at the +1/+2 sites and conformationalchange of loops

The ligand-bound structure obtained from the pNPG3-

soaked crystal contained one �-cellobiose molecule, which

occupies subsites +1 and +2 in the substrate-binding cleft

(Fig. 2). The observed �-cellobiose is considered to be a

product of the hydrolysis of pNPG3 under the crystallization

conditions at pH 5. This is plausible because the GH family 6

enzymes perform hydrolysis of �-1,4-glycosidic bonds with

inversion of anomeric configuration, and are known to cleave

chromophoric cello-oligosaccharides to produce cellobiose

units (Claeyssens et al., 1989). The cleavage pattern of pNPG3

in this structure is different from the CcCel6A structure, in

research communications

400 Tachioka et al. � Family 6 cellobiohydrolase Acta Cryst. (2017). F73, 398–403

Figure 1Overall structure of PcCel6A with cellobiose bound at subsites +1 and +2. The N-terminal and C-terminal loops covering the catalytic centre arecoloured green and cyan, respectively.

Figure 2Close-up view of subsites +1 and +2 of PcCel6A with cellobiose as a ligand. The |Fo|� |Fc| map was calculated without ligand atoms and contoured at the3� level. The Tris molecule at subsite �1 was modelled in a double conformation.

which nonhydrolyzed pNPG3 bound at subsites +1 to +4

(PDB entry 3voi; Tamura et al., 2012), and from the CcCel6C

structure, in which two pNPG2 molecules bind at subsites �3

to �1 and +1 to +3 (PDB entry 3abx; Liu et al., 2010).

As in other GH family 6 enzymes, the �1 subsite is often

occupied by molecules other than waters and sugars, such as

cations and low-molecular-weight compounds (Supplemen-

tary Table S1), and an unmodelled electron-density blob was

similarly observed at the �1 site in the present liganded

structure. An octahedrally coordinated Mg2+ ion found at

subsite �1 of the CcCel6C–pNPG3 structure (PDB entry

3voi), where the ligands only occupied plus-numbered subsites

as in this PcCel6A–cellobiose complex, led us to initially place

a hydrated Ca2+ ion into the density blob. However, the B

factor of the hydrated Ca2+ ion (48.8 A2) was quite high

compared with those of cellobiose (18.5–26.8 A2) and the

water molecules around the subsite (approximately 15–30 A2).

The automated ligand-identification tool in PHENIX

(Terwilliger et al., 2006) predicted Tris as the most favourable

of the molecules in the crystallization condition, and the B

factors for a single Tris molecule were in a reasonable range

(29.1–33.4 A2). The second conformation of the Tris molecule

was placed manually into the residual electron density of this

model, and a molecule with multiple conformations appears to

be the most probable explanation, with B factors of 18.3–

24.8 A2. The Tris molecule made hydrogen bonds to O4 of

glucose at subsite +1, the OD1 atom of Tyr164, the NZ atom of

Lys388, the carbonyl O atom and the OD1 atom of Asp394 in

addition to several water molecules. No water molecule

corresponding to the catalytic water was found in the structure

because of the binding of the putative Tris molecule.

The tunnel-enclosing N- and C-terminal loops of GH family

6 enzymes have been observed in several conformations in

previous crystallographic studies (reviewed in Payne et al.,

2015). Two major conformations of the loops are known, an

‘open’ conformation that makes the subsites more accessible

and a ‘closed’ conformation for arrangement of the catalytic

site residues into the catalytically competent configuration. In

the cellobiose-bound structure, the loops adopted the closed

conformation, in contrast to the open conformation observed

in the apo structure. As summarized in Supplementary Table

S1, the loops may adopt either conformation in enzymes for

which the structures are known. The occupation of +1 and +2

sites by sugar moieties results in the closure of the loops, with

the exception that proton-accepting residues are ‘off’ confor-

mation and interact with serines on the loop, as in HiCel6A

structures where sugars bind in upside-down configurations

(PDB entries 1oc5, 1oc7 and 1ocj; Varrot et al., 2003).

3.3. Noncatalytic but notable residues of PcCel6A

Mutagenesis work performed on PcCel6A revealed several

properties of noncatalytic residues of this enzyme (see Hein-

zelman et al., 2009; Ito et al., 2013; Tachioka et al., 2016). In our

recent work, the substitution mutation W267C was found to be

a critical mutation for the degradation of crystalline cellulose

IIII but not of amorphous phosphoric acid-swollen cellulose

(Tachioka et al., 2016), which was a similar finding to the work

on TrCel6A (Koivula et al., 1998). Trp267 is located at the

entrance to the catalytic site at subsite +4 and probably plays a

specialized role in the recognition and recruitment of a single

cellulose chain from the crystalline surface (Fig. 3a). Hein-

zelman and coworkers successfully engineered PcCel6A to

improve its thermostability by a remarkable 10�C by a single

mutation of Cys393 to serine (Heinzelman et al., 2009). This

free cysteine makes a weaker hydrogen bond to the carbonyl

of Pro418, with a distance of about 3.3 A (Fig. 3b), than the

serine residues that are found in other homologous fungal

structures which originally have a serine instead of cysteine

(2.6–2.7 A in Chaetomium thermophilum Cel6A and

CcCel6C), which is consistent with their discussion. Ito and

coworkers also improved the thermostability of PcCel6A by

16 cumulative mutations (Ito et al., 2013). The largest stabili-

zation of 1.2�C was achieved by a mutation of Met257 to

isoleucine, and the authors deduced from modelling PcCel6A

that an exchange to more hydrophobic residues stabilizes the

research communications

Acta Cryst. (2017). F73, 398–403 Tachioka et al. � Family 6 cellobiohydrolase 401

Figure 3Enlarged views of Trp267 (a), Cys393 (b) and Met257 (c). Each residue is coloured magenta. The red–white colour in (c) represents the hydrophobicityof the residues.

structure in its buried protein environment. Our results reveal

that Met257 is surrounded by hydrophobic side chains of

�-helix and supports their findings (Fig. 3c).

3.4. Loop flexibility

The flexibility and mobility of the loops are important for

the catalytic activity. As noted above in x3.2, the loops of

PcCel6A remained open in the absence of bound ligand, but

occupation of the substrate-binding sites induced conforma-

tional change of the loops, resulting in a narrowing of the

active-site cleft. To further examine the situation, the relative

B factors of residues and r.m.s.d. values among the observed

conformations of PcCel6A were visualized by means of a

blue–red colour scale and worm representations, respectively

(Fig. 4). The residues in the loops have consistently high B

factors compared with other regions of the protein. As shown

in Fig. 4(a), the N- and C-terminal loops of the PcCel6A apo

structure were both modelled as double conformations with

similar occupancies ranging from 0.46 to 0.54. In Figs. 4(b) and

4(c), the relatively open and most open apo structures are

compared with the cellobiose-bound structure, respectively.

The most dynamic conformational change occurred in the

N-terminal loop concomitantly with ligand binding, as shown

by a ribbon representation, the thickness of which is in

proportion to the r.m.s.d values between two conformations.

In the C-terminal loop region, on the other hand, the r.m.s.d.

values between the apo and cellobiose-bound structures were

similar to that between the two conformations of the apo

structure. Therefore, both loops showed clear flexibility, but

their mobility and response to ligand binding were different.

Although the role of each loop in GH family 6 cellobio-

hydrolases is not fully understood, it has been suggested that a

conformational change of the loops would occur during the

catalytic cycle (see Payne et al., 2015). In this study, we

determined both the open and closed structures of the CD of

PcCel6A. Careful analysis of the flexibility and mobility of the

N-terminal and C-terminal tunnel-enclosing loops revealed

differences in their characteristics, and we speculate that the

loops are involved in both the chemical hydrolysis reaction of

a single polysaccharide chain and the physical interaction with

the crystalline surface of cellulose. The present structural

information should afford insight into the structure–function

relationship and provide a basis for more detailed mechanistic

studies of this industrially and scientifically important enzyme.

Acknowledgements

We thank Dr Shinya Fushinobu for the provision of beam time

and all beamline staff members at the Photon Factory for their

kind help with data collection. We are also grateful to Dr Jerry

Stahlberg at the Swedish University of Agricultural Sciences

for valuable discussions and suggestions. This work was

performed under the approval of the Photon Factory Program

Advisory Committee (Proposal Nos. 2013G191 and

2013G194).

Funding information

This research was supported by a Grant-in-Aid for Innovative

Areas from the Japanese Ministry of Education, Culture,

Sports and Technology (MEXT) to KI (Nos. 24114001 and

24114008), a Grant-in-Aid for Scientific Research from the

Japan Society for the Promotion of Science (JSPS) to KI (B,

No. 24380089), an Impulsing Paradigm Change through

Disruptive Technologies (ImPACT) from the Japan Science

and Technology Agency (JST) to KI and a Grant-in-Aid for

JSPS Fellows from JSPS to MT (15J10657).

References

Adams, P. D. et al. (2010). Acta Cryst. D66, 213–221.

research communications

402 Tachioka et al. � Family 6 cellobiohydrolase Acta Cryst. (2017). F73, 398–403

Figure 4Superposition of the apo and cellobiose-bound structures of PcCel6A. (a) The apo structure was observed in multiple conformations: the ‘relativelyopen’ and ‘most open’ conformations are shown. (b) Superposition of the relatively open apo structure and the cellobiose-bound structure. (c)Superposition of the most open apo structure and the cellobiose-bound structure. The width of the ribbon in each figure represents the r.m.s.d. valuebetween the superposed conformations. The relative B factors were calculated for each apo and cellobiose-bound structure.

Claeyssens, M., Van Tilbeurgh, H., Tomme, P., Wood, T. M. & McRae,S. I. (1989). Biochem. J. 261, 819–825.

Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. (2010). ActaCryst. D66, 486–501.

Heinzelman, P., Snow, C. D., Smith, M. A., Yu, X., Kannan, A.,Boulware, K., Villalobos, A., Govindarajan, S., Minshull, J. &Arnold, F. H. (2009). J. Biol. Chem. 284, 26229–26233.

Holm, L. & Rosenstrom, P. (2010). Nucleic Acids Res. 38, W545–W549.

Igarashi, K., Maruyama, M., Nakamura, A., Ishida, T., Wada, M. &Samejima, M. (2012). J. Appl. Glycosci. 59, 105–110.

Ito, Y., Ikeuchi, A. & Imamura, C. (2013). Protein Eng. Des. Sel. 26,73–79.

Kelley, L. A. & Sternberg, M. J. E. (2009). Nature Protoc. 4, 363–371.

Koivula, A., Kinnari, T., Harjunpaa, V., Ruohonen, L., Teleman, A.,Drakenberg, T., Rouvinen, J., Jones, T. A. & Teeri, T. T. (1998).FEBS Lett. 429, 341–346.

Koivula, A. et al. (2002). J. Am. Chem. Soc. 124, 10015–10024.Liu, Y., Yoshida, M., Kurakata, Y., Miyazaki, T., Igarashi, K.,

Samejima, M., Fukuda, K., Nishikawa, A. & Tonozuka, T. (2010).FEBS J. 277, 1532–1542.

Mertz, B., Kuczenski, R. S., Larsen, R. T., Hill, A. D. & Reilly, P. J.(2005). Biopolymers, 79, 197–206.

Otwinowski, Z. & Minor, W. (1997). Methods Enzymol. 276, 307–326.Payne, C. M., Knott, B. C., Mayes, H. B., Hansson, H., Himmel, M. E.,

Sandgren, M., Stahlberg, J. & Beckham, G. T. (2015). Chem. Rev.115, 1308–1448.

Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S.,Greenblatt, D. M., Meng, E. C. & Ferrin, T. E. (2004). J. Comput.Chem. 25, 1605–1612.

Rouvinen, J., Bergfors, T., Teeri, T., Knowles, J. K. C. & Jones, T. A.(1990). Science, 249, 380–386.

Spezio, M., Wilson, D. B. & Karplus, P. A. (1993). Biochemistry, 32,9906–9916.

Tachioka, M., Sugimoto, N., Nakamura, A., Sunagawa, N., Ishida, T.,Uchiyama, T., Igarashi, K. & Samejima, M. (2016). Biotechnol.Biofuels, 9, 1–10.

Tamura, M., Miyazaki, T., Tanaka, Y., Yoshida, M., Nishikawa, A. &Tonozuka, T. (2012). FEBS J. 279, 1871–1882.

Tempelaars, C. A. M., Birch, P. R. J., Sims, P. F. G. & Broda, P. (1994).Appl. Environ. Microbiol. 60, 4387–4393.

Terwilliger, T. C., Klei, H., Adams, P. D., Moriarty, N. W. & Cohn, J. D.(2006). Acta Cryst. D62, 915–922.

Varrot, A., Frandsen, T. P., von Ossowski, I., Boyer, V., Cottaz, S.,Driguez, H., Schulein, M. & Davies, G. J. (2003). Structure, 11,855–864.

Winn, M. D. et al. (2011). Acta Cryst. D67, 235–242.

research communications

Acta Cryst. (2017). F73, 398–403 Tachioka et al. � Family 6 cellobiohydrolase 403